Deploying packages to KAPPA-Automate#

You may have noticed it but when you are in the Script tab, there is a Package selector, it allows you to select packages that you need to run your user task. For the GOR user task we did not need any packages to run the user task, but it could happen that you need to use some packages like numpy, pandas or scipy. In this section we will briefly cover working with packages in KAPPA-Automate.

Navigate to the Automation section, select the User Tasks option on the left and then click the Packages section. You can see that there is at least one package in the list (the built-in Computation and Math Libraries (Linux)) that contains a few standard python libraries like numpy, pandas, scipy, scikit-learn.

If you need to upload a custom package:

  • Click on the Create button

  • Give an explicit name to your package

  • Select the scripting language, in our case it will be Python

  • You can give a description to your package

  • And, finally, you just need to provide a zip file with the package content

Once you have deployed a package, you can later update it, if needed:

  • Double click on the package to be updated or click on it and select the Edit button.

  • It is possible to change the name, the description and the content of the package. Tick the box update file and provide a new zip of the package to change the content of the deployed package.

Note

You cannot update built-in packages. So for example, if you want to use a different version of numpy package you will need to create and publish a new KAPPA-Automate package with this specific version.

Important

Don’t forget that when you deploy a package, you also need to deploy all the dependencies of the package. To make it easier, it is possible to have multiple python packages in a single KAPPA-Automate package.

Warning

Your Python package must be compatible with Linux Systems since the user task environment is running on Linux. To help you we develop a tool, see KAT tool